ClawHub
Back to skill

Security audit

Minimum Viable Product

Security checks for vulnerabilities and agentic risk

Overview

This product-planning skill is proportionate to its purpose, though its advice to omit login features needs care when testing with real customers.

Treat this as planning advice. Keep authentication and authorization wherever private customer information or privileged functions require them, and use hosted checkout for payment tests. The package does not itself deploy a service or collect customer data.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

The workflow directs removal of authentication without warning about protecting participant data or restricting access.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

The instruction to strip authentication and accounts unless the test requires them frames access controls as experimental features rather than safeguards required by data sensitivity and exposure. Because the skill also recommends live paid pilots, email collection, and pre-orders, following this guidance could produce an externally accessible MVP that exposes participant information or privileged functionality. The conditional wording reduces the risk, but does not explicitly preserve security requirements independently of the assumption being tested.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.